Fighting cancer: CSAIL scientists’ AI algorithms can detect dangerous disease

In the United States, researchers have created AI algorithms that can detect pancreatic cancer at early stages, reports Engadget.

The Computer Science and Artificial Intelligence Laboratory (CSAIL) of the Massachusetts Institute of Technology (MIT) is responsible for the development. Scientists have combined the algorithms into a neural network called PRISM.

The aim of the development is to detect pancreatic ductal adenocarcinoma (PDAC), which is called the most common type of pancreatic cancer.

Modern diagnostic standards allow detecting the disease in only 10% of cases among patients examined by specialists. The new method proved to be more efficient, with AI boosting this figure to 35%.

The neural network was programmed based on access to various sets of real electronic medical records in medical institutions across the United States. More than 5 million electronic medical records of patients have been uploaded to it.

PRISM uses clinical and laboratory data to work. To make a diagnosis, the neural network takes into account demographic data, previous diagnoses, medications, test results, age, and risk factors.

Earlier it became known that in the United States, scientists have discovered a new way to destroy cancer cells. During the research, they directed near-infrared light at the aminocyanine molecules, which made them vibrate. This was enough to break the membranes of cancer cells.